Systems Engineer deployed onsite at a UAE government agency, running their OpenShift, Linux, Solaris, and enterprise storage environment in a largely air-gapped environment. On the AI side, I build and run a fully local GenAI stack on self-managed Kubernetes.
- Model serving: vLLM · TGI · Ollama
- Models: Qwen3.5 · DeepSeek · LLaMA
- Quantization: AWQ · GPTQ · GGUF
- RAG: LangChain · Chroma · nomic-embed-text
- Agents: LangGraph (stateful workflows, tool selection, conditional edges, handoffs)
- Orchestration: Kubernetes · Docker · Red Hat OpenShift · Nutanix
- Observability: Elasticsearch · Grafana
- IaC & CI/CD: Terraform · Ansible · GitHub Actions
- OS: Linux · Solaris
- Languages: Python · Bash · SQL
- Local LLM serving stack — vLLM/TGI/Ollama on Kubernetes with quantized open-weight models
- LangGraph agent orchestration + RAG pipeline over a personal Obsidian knowledge base
| Project | Description |
|---|---|
| Advertisement Detection | NLP classifier on 1M+ row URL dataset · 80% accuracy · TensorFlow / XGBoost / LightGBM |
| Amazon Sentiment Analysis | 227k reviews · 90% accuracy · NLP + neural networks |
| Twitter Sentiment Analysis | 10k+ tweets · 92% accuracy · NLP + ensemble methods |
| vscan | Python security scanner · multi-API · published on PyPi |
omar@itsmokha.com · itsmokha.com · linkedin.com/in/m-omarkhan